As optical signals used in optical communications carry ever increasing data rates according to an ever widening variety of data standards, it becomes desirable to provide switching at the photonic level, i.e., without resorting to electronic circuitry for converting the optical signals into the electrical domain before switching is performed. These types of optical switches are referred to as photonic (or OOO-short for “Optical Input, Optically Switched, Optical Output”) switches.
The desirable characteristics of a photonic switch are scalability, robustness and the ability to provide non-blocking performance in a compact low-cost package. Generally speaking, first-generation photonic switches afford at most two of these benefits at the expense of the other(s) in packages compromised in size and cost due to the complex, usually fiber-guided, interconnect between the various modules of the switch.
For example, first-generation photonic switches that are scalable by virtue of a modular design (e.g., multiple planes on a per-wavelength, or per-wavelength-group, basis) typically require a wavelength conversion unit to provide a satisfactory level of residual blocking performance. This introduces inefficiencies in provisioning the switch. Also, since optical signals are converted into the electrical domain for the purposes of wavelength conversion, switches of this type lose the designation of being truly photonic in nature. Moreover, in lambda-plane switches, the optical interconnect requires up to thousands of individual optical fiber connections, which can be reduced in size somewhat by the provision of an orthogonal shuffle function, but this nevertheless results in a non-compact solution.
Other designs, such as multi-stage photonic switches (e.g., CLOS), can be made non-blocking through dilation or path rearranging, but do not scale well to accommodate an increase in the number of input signals. In particular, the complexity of the interconnect between stages becomes intractable as the number of input signals increases. Furthermore, in addition to introducing a delay, the multi-stage characteristic of these switches imparts a higher path loss due to multiple lossy switching operations in series that need to be compensated for in the design.
Still other first-generation photonic switch architectures, such as the Xros X-1000, utilize opposing arrays of independently controllable mirrors at the end of an optical chamber to achieve non-blocking performance. However, these switches tend to be large in size, have low tolerance to manufacturing error and also do not scale well due to a lack of modularity. In addition, such switches have a complex fiber-based interconnect.
Against this background, it is clear that there exists a need in the industry for improvement in the area of photonic switches.